AI isn't the future - it's the now!

Discover the key insights from the 2024 GCSG presentation on leveraging Artificial Intelligence to optimize daily tasks.

Key Takeaways from the 2024 GCSG Presentation

The clinical supply world gathered in Naples, FL last week for the annual Global Clinical Supplies Group conference (GCSG) and by all accounts it was a huge success.  Perhaps the most educational event of the annual conference calendar, this year's conference was packed with new thoughts, regulations, applications, suppliers, and networking.  I was fortunate to have an opportunity to provide a workshop this year on Artificial Intelligence and its ability to streamline your daily activities.  

What are attendees asking for from AI applications

We started the presentation by asking the group for their ideas or wants from AI in their daily tasks.  The responses provided a mix of clinical supply related tasks, IRT/RTSM workflows, and personal organizational needs.  While going into detail on each of these ideas may be a topic of future blog posts, for today we're happy to provide a list of ideas for any developers or hobbyists to delve into.

• Forecasting - to replace manual excel workbooks
• Bots to answer standard and routine questions that customers ask, like why does they IRT say I have no drug on site to dispense the visit, but there is a shipper of IP right here.  (Hint, you need to run the "shipment acknowledgment at site" transaction)
• Automation of manual data entry for packaging and labeling
• Oversight and enforcement of standards across vendors
• Prioritization of personal tasks, schedules, emails, etc...
• Check for available site stock against upcoming visits
• General data recall from  protocols or requirements documents

Importance of Data Privacy in AI Applications

Before we dove too deeply into the use of AI and how to leverage it's power, we stressed the importance of either remaining in compliance with your data privacy policy, or defining one if there isn't a policy in place already.  This will evolve in the next weeks and months for sure, particular with the new EU Artificial Intelligence Act taking effect this spring.  

• Data Minimization: Collect only the data necessary for the task at hand.
• Transparency: Clearly inform users about what data is collected, how it's used, and who it's shared with.
• Secure Storage & Transfer: Implement robust security measures to protect data at rest and in transit.
• Access Controls: Restrict access to data to only those who need it for legitimate purposes.

Adherence to Company Data Policies

• Policy Overview: Summarize the key aspects of your company’s data privacy policies.
• Compliance Training: Regular training sessions for employees on data privacy standards and protocols.
• Regular Audits: Conduct periodic audits to ensure compliance with data privacy policies and regulations.
• Incident Response Plan: Establish and rehearse a plan for responding to data breaches or other privacy incidents.

To enhance the efficiency of using AI, consider the following tools:

One of my favorite areas to speak on is actually one level higher than implementation of AI models.  Another way to put this is that preparing data and understanding how to use the AI tools is just as critical as knowing which AI tools to use.  As an example, I outlined a few tools that can be leveraged as a sort of data intermediary to better leverage AI.

When copying and paste   thete theting data into a generative AI tool, like ChatGPT for example, there are some formats that it can understand better than others.  Admittedly, I've come to this conclusion pragmatically simply by trial and error, but in general I've had good results.  I've leveraged a language called Markdown (MD) for pasting table data from excel.  MD is markup language just like HTML but is natural and human readable.  I've leveraged tools like Table-to-Markdown to easily convert data into a format that AI can easily and more effectively process.  While this is just one example of tools that are available, and they're so many, it's an important to think about what we put into an AI model to better understand what we will get back.  Like all things, garbage in - garbage out.  

Next Steps in AI Technology

If you consider yourself a power user of AI tools, then you don't need this post to know how exciting a time we are living in.  While there is legitimate reason for some to be skeptical and even afraid of the power of AI, I've seen far more reasons to be curious and energized by the potential.  Building AI tools and apps is actually easier than you may have thought and it's becoming more and more accessible.  

Some tools that I've recently started to explore in more depth combine the idea of low-code/no-code with the power of AI.  Many of these platforms require some sort of account or subscription but there are also free tools available.  My current tool box of power-user AI resources includes Flowise, Zapier, and Power Automate.  There is a distinct difference between automation and Artificial Intelligence, but these tools allow you to quickly and easily arrive at their intersection to start creating real-world solutions.